SYMPOSIA PAPER Published: 01 January 1998

Assessment of Selenium Food Chain Transfer and Critical Exposure Factors for Avian Wildlife Species: Need for Site-Specific Data


Observations of selenium poisoning in Belews Lake, NC in the mid-1970s and Kesterson Reservoir, CA in the mid-1980s precipitated a large number of selenium studies. Numerous authors have evaluated the potential for selenium to cause ecologically significant effects via food chain transfer in aquatic ecosystems, especially wetlands. Additionally, bioaccumulation models have been proposed for estimating selenium concentrations in food chains and water that should not be exceeded in order to avoid reproductive effects in avian and aquatic species. The current national chronic ambient water quality criterion (WQC) for protection of aquatic life is 5 μg/L. Scientists with the U.S. Fish and Wildlife Service have recommended setting the ambient water quality criterion at 2 μg/L for both aquatic and wildlife protection.

Reported site-specific variations in selenium's effects on aquatic life and birds prompted us to re-evaluate the basis for the 2 μg/L recommendation, and in particular one of the wildlife bioaccumulation models used to support this value. We used a probabilistic approach to assess water, food chain, and bird egg residues from 15 sites. Our data evaluation indicates significant differences in selenium accumulation in invertebrates and bird eggs among sites and among species. Both a two-step regression model (water → food chain → bird eggs) and a one-step regression going directly from waterborne selenium (WS) to mean egg selenium (MES) were fitted to all data for 15 sites and four bird species. The one-step model contained less variability than the two- step model and had a coefficient of variation (r2) of 0.67. Uncertainty analysis of the regression models provided a distribution of waterborne selenium concentrations associated with bird egg tissue residues. Using the 10th and 50th percentiles of these distributions, we calculated waterborne selenium concentrations between 6.8 and 46 μg/L that are protective of birds. These values are associated with an effects threshold of 20 mg/kg selenium dry weight in bird eggs, which is the EC10 for mallard duck embryo teratogenis (Skorupa et al. 1996). The 10th percentile of this distribution, 6.8 μg/L, is slightly above the EPA water quality criterion of 5 μg/L. The water concentrations protective of birds range from slightly more than the current EPA WQC (6.8 vs. 5.0 μg/L) to a factor of 10 or greater at some sites.

Our results also indicate a reasonably strong correlation between water and mean egg selenium concentrations. However, site-specific factors strongly influence this relationship, and when waterborne selenium approaches or exceeds the WQC, collection of site-specific data would be appropriate to accurately assess the WS to MES relationship. Evaluation of the site-specific relationship between WS and MES can determine whether site-specific differences are important and whether or not the EPA WQC is likely to be over protective of bird populations. In this paper, we question the need for the WQC to be set at 2 μg/L to protect aquatic birds. Overall, bird-egg residues appear to be the best tool for assessing potential for risk to birds from selenium.

Author Information

Adams, WJ
Kennecott Utah Copper, Magna, UT
Brix, KV
Parametrix, Inc., Kirkland, WA
Cothern, KA
Parametrix, Inc., Kirkland, WA
Tear, LM
Parametrix, Inc., Kirkland, WA
Cardwell, RD
Parametrix, Inc., Kirkland, WA
Fairbrother, A
ecological planning and toxicology, inc., Corvallis, OR
Toll, JE
Parametrix, Inc., Kirkland, WA
Price: $25.00
Contact Sales
Reprints and Permissions
Reprints and copyright permissions can be requested through the
Copyright Clearance Center
Developed by Committee: E47
Pages: 312–342
DOI: 10.1520/STP12173S
ISBN-EB: 978-0-8031-5384-4
ISBN-13: 978-0-8031-1485-2